Today I would like to add something to my answer sent two days ago; I am sending this to the whole List because I hope that it may be of interest to cochlear modelers. Last October I presented my own cochlear model to the annual meeting of the Canadian Acoustical Association in Niagara-on-the Lake; see R. F., "Old and New Cochlear Maps", Canadian Acoustics Vol. 37 No. 3 (2009) 174-175. That model involves two cochlear resonators, namely the "IOCR" [internal organ-of-Corti resonator; springs = outer hair cells (OHCs) and surrounding structures] and the "BMR" (basilar-membrane resonator; springs = fibres of the BM). At given distance-from-base in the basal half, the resonance frequency of the local IOCR is lower, by about one octave, than that of the local BMR. During a sine-tone, the IOCR enables the OHCs to "actively" feed energy into the travelling wave, to thus give rise to the "active" excitation peak (prominent at low sound level), and to make the real (i.e., resistive) part of the low-sound-level BM impedance negative. That model has been found to be adequate by showing that it is consistent with five experimental sine-tone functions of distance-from-base, namely (1) the low-level BM oscillation velocity, (2) the high-level (or any-level post-mortem) BM oscillation velocity, (3) the corresponding BM-oscillation phase (which depends weakly only on level), (4) the real part of the BM impedance, and (5) the imaginary part of the BM impedance. The experimental data are briefly described in my above-mentioned paper.
Datum: 23.05.2010 04:18
Betreff: Re: comparing cochlear models ?
Thanks a lot for the reference and I appreciate all the other usefull comments too.
I like the cochlear model of Paul Kolston, described in his paper "The importance of phase data and model dimensionality to cochlear mechanics", Hear. Res. 145 (2000) 25-36. In his Fig. 8, Kolston proves that his model is good by showing that it correctly predicts two experimental functions of distance-from-base: 1) the BM velocity at given sine-tone frequency, and 2) the corresponding phase. The agreement of such theoretical curves with experimental data can be quantified by chi-squared, i.e. by the sum of
[(f_theo-f_exper)/df_exper]^2, taken over all experimental values of the function f_exper; df_exper is the experimental uncertainty of f_exper.
With best wishes,
Dr. phil. nat.,
r. PSI and ETH Zurich,
Phone: 0041 56 441 77 72.
Mobile: 0041 79 754 30 32.
E-mail: reinifrosch@xxxxxxxxxx .